Depression is among the most prevalent of all psychiatric disorders, accounting for over 20 percent of economic costs for all mental illness. A great deal of theoretical attention has focused on the possibility that negative thinking might represent not only a feature of depression, but a vulnerability factor for this disorder as well. A recent influential research paradigm has operationalized depressotypic cognitions in terms of selective attention to, and memory for, negative emotional stimuli. The overall goal of the proposed investigations is to utilize this paradigm to investigate the role of cognitive biases in onset and course of depression, and to examine neurobiological foundations of cognitive biases in depression. Specific aims include (1) examining the utility of cognitive biases to predict the course of depressive symptoms and diagnostic status over a two-year period; (2) localizing the neurobiological underpinnings of these biases; and (3) examining the breadth of these biases and their specificity to depression. To achieve these aims, standardized cognitive information- processing tasks will be used to identify 30 "high-bias" and 30 "low- bias" depressed patients in psychiatric outpatient clinics. The nature and breadth of these biases in the depressed patients will be compared to cognitive biases among 30 patients diagnoses with generalized anxiety disorder, 30 patients diagnosed with social phobia, and 30 non-patient controls. Each depressed patient will be followed for one year, and the degree of cognitive biases will be reassessed when the patient achieves clinical remission. Hypotheses concerning the neurobiological underpinnings of depressotypic cognitive biases will be tested by conducting functional magnetic resonance imaging (MRI) of depressed (and later, remitted) patients while they are performing information- processing tasks.